Information processing system

Abstract

Provided is an information processing system which excels in convenience and usefulness. A pair of an original sentence and a translated sentence is acquired, and is divided into phrases using a morpheme analyzer. The divided phrases are converted into vectors using an embedding model. A degree of similarity is calculated by using the vectors, and a similarity matrix is created. The similarity matrix is normalized to create a normalized matrix. A probability is calculated from the normalized matrix. A correspondence table listing a combination of an original sentence phrase and a translated sentence phrase is created, and a non-correspondence list listing phrases not included in the correspondence table is created. A prompt which includes the original sentence, the translated sentence, the correspondence table, and the non-correspondence list is created. A response to the prompt is generated using a large-scale language model, and a bilingual dictionary is created on the basis of the response results. A translated sentence phrase candidate and a determination reason are output by referring to the bilingual dictionary.

Description

Information Processing Systems

[0001] One aspect of the present invention relates to an information processing system and an information processing method.

[0002] Note that one embodiment of the present invention is not limited to the above technical field. The technical field of one embodiment of the invention disclosed in this specification and the like relates to an object, a method, or a manufacturing method. Alternatively, one embodiment of the present invention relates to a process, a machine, manufacture, or a composition of matter. Therefore, more specifically, examples of the technical field of one embodiment of the present invention disclosed in this specification include a semiconductor device, a display device, a light-emitting device, a power storage device, a memory device, a driving method thereof, or a manufacturing method thereof.

[0003] Research and development of computer-based machine translation is currently underway. Examples of machine translation include rule-based machine translation, statistical machine translation that uses language models and translation models, and neural machine translation that uses artificial neural networks (ANNs, hereinafter referred to as neural networks).

[0004] In particular, in recent years, the development of language models using neural networks has been actively carried out, and large-scale language models (LLMs) have been attracting particular attention. A large-scale language model is a natural language processing model trained using a large amount of data. A large-scale language model can realize a dialogue model that responds to user instructions, for example. Non-Patent Document 1 discloses GPT-4 (registered trademark) (Generative Pre-trained Transformer 4) as a large-scale language model, and also discloses ChatGPT as a chat service using a large-scale language model.

[0005] Summary of ChatGPT / GPT-4 Research and Perspective Towards the Future of Large Language Models, Yiheng Liu et al. (Submitted on 4 Apr 2023, [online], Internet <URL: https: / / arxiv.org / abs / 2304.01852>

[0006] The accuracy of machine translation is improving year by year, and users can use machine-translated text as is in personal documents, etc., as long as the content is conveyed to the reader.

[0007] However, there are many documents, such as contracts, official documents, patent documents, and academic papers, where errors are unacceptable. When translating such documents, whether manual or machine translation is used, checking and correcting the text after translation is important. Note that manual checking and correction of machine-translated text is sometimes called post-editing, and the international standard ISO 18587 specifies requirements for post-editing.

[0008] One method for checking and correcting translated text involves manually building a bilingual dictionary and checking it, but this method has issues such as the time required to build the dictionary and its vulnerability to spelling variations. Furthermore, when translating using a bilingual dictionary, it can be difficult to translate words that are not registered in the dictionary. Furthermore, while using a large-scale language model allows for flexible checking, there are issues with accuracy and concerns that checks may be missed.

[0009] An object of one embodiment of the present invention is to provide an information processing system that supports checking and correcting a translated sentence.An object of one embodiment of the present invention is to provide a novel information processing system that is highly convenient or useful.An object of one embodiment of the present invention is to provide an information processing method that supports checking and correcting a translated sentence.An object of one embodiment of the present invention is to provide a novel information processing method that is highly convenient or useful.An object of one embodiment of the present invention is to provide a novel information processing system, a novel information processing method, or a novel semiconductor device.

[0010] Note that the description of these problems does not preclude the existence of other problems. Note that one embodiment of the present invention does not necessarily solve all of these problems. Note that problems other than these will become apparent from the description of the specification, drawings, claims, etc., and it is possible to extract other problems from the description of the specification, drawings, claims, etc.

[0011] One aspect of the present invention includes a first information processing device and a second information processing device, and the first information processing device has first to eighth functions. The first function is a function of acquiring a first sentence written in a first language and a second sentence obtained by translating the first sentence into a second language. The second function is a function of dividing the first sentence into a plurality of first phrases. The third function is a function of dividing the second sentence into a plurality of second phrases. The fourth function is a function of calculating the probability that one of the plurality of first phrases and one of the plurality of second phrases are translations of each other, and creating a matrix. The fifth function is a function of calculating the probability that one of the plurality of first phrases and one of the plurality of second phrases are translations of each other, and creating a matrix. The sixth function is a function of creating a prompt including a first sentence, a second sentence, the correspondence table, and the unmatched list; the seventh function is a function of sending the prompt to a second information processing device; and the eighth function is a function of acquiring, as a response result to the prompt, phrases included in the unmatched list, phrases determined to be translations of the phrases, and the reason for the judgment from the second information processing device; the second information processing device has the function of acquiring the prompt and generating a response result using a large-scale language model.

[0012] Alternatively, one embodiment of the present invention includes a first information processing device, a second information processing device, and a third information processing device. The first information processing device has first to twelfth functions. The first function is a function of acquiring a first sentence written in a first language and a second sentence obtained by translating the first sentence into a second language. The second function is a function of dividing the first sentence into a plurality of first phrases. The third function is a function of dividing the second sentence into a plurality of second phrases. The ninth function is a function of dividing the plurality of first phrases and the plurality of second phrases. a function of transmitting a plurality of first vectors corresponding to the plurality of first phrases and a plurality of second vectors corresponding to the plurality of second phrases to a third information processing device; a tenth function of acquiring from the third information processing device a plurality of first vectors corresponding to the plurality of first phrases and a plurality of second vectors corresponding to the plurality of second phrases; an eleventh function of calculating a similarity between the first vector and the second vector and creating a similarity matrix including the similarity; a twelfth function of normalizing the similarity matrix and creating a normalized matrix; and a fourth function of extracting a plurality of first vectors corresponding to the plurality of first phrases and a plurality of second vectors corresponding to the plurality of second phrases from the normalized matrix. the fifth function is a function of creating a correspondence table listing combinations of first phrases and second phrases whose probability is higher than a threshold, and an unmatched list listing phrases not included in the correspondence table; the sixth function is a function of creating a prompt including a first sentence, a second sentence, the correspondence table, and the unmatched list; the seventh function is a function of sending the prompt to a second information processing device; and the eighth function is a function of acquiring from the second information processing device, as a response result to the prompt, phrases included in the unmatched list, phrases determined to be translations of the phrases, and reasons for the judgment; the second information processing device has a function of acquiring the prompt and generating a response result using a large-scale language model; and the third information processing device has a function of converting the first phrase into a first vector and the second phrase into a second vector by processing using an embedding model.

[0013] In the sixth function, it is preferable that the prompt be created by inserting the first sentence, the second sentence, the correspondence table, and the uncorresponding list into a template.

[0014] It is preferable that the first information processing device has a function of creating a bilingual dictionary that stores phrases included in the unsupported list, phrases determined to be translations of the phrases, and reasons for the determination, based on the response results.

[0015] It is preferable that the first information processing device has a function of referring to a bilingual dictionary and, if there are multiple second phrases that are candidate translations for one of multiple first phrases, outputting the candidate second phrases and the reason for the determination.

[0016] According to one embodiment of the present invention, an information processing system that supports checking and correcting a translated sentence can be provided. According to one embodiment of the present invention, a novel information processing system with excellent convenience or usefulness can be provided. According to one embodiment of the present invention, an information processing method that supports checking and correcting a translated sentence can be provided. According to one embodiment of the present invention, a novel information processing method with excellent convenience or usefulness can be provided. According to one embodiment of the present invention, a novel information processing system, a novel information processing method, or a novel semiconductor device can be provided.

[0017] Note that the description of these effects does not preclude the existence of other effects. Note that one embodiment of the present invention does not necessarily have all of these effects. Note that effects other than these will become apparent from the description in the specification, drawings, claims, etc., and it is possible to extract other effects from the description in the specification, drawings, claims, etc.

[0018] FIG. 1 is a diagram illustrating an example of the configuration of an information processing system. FIG. 2 is a diagram illustrating an example of the configuration of an information processing device. FIG. 3 is a flowchart illustrating an example of an information processing method. FIGS. 4A to 4C are flowcharts illustrating an example of an information processing method. FIG. 5 is a diagram illustrating an example of the configuration of an information processing system. FIGS. 6A to 6D are diagrams illustrating an example of data related to the operation of the information processing system. FIGS. 7A and 7B are diagrams illustrating an example of data related to the operation of the information processing system. FIGS. 8A and 8B are diagrams illustrating an example of data related to the operation of the information processing system. FIG. 9 is a diagram illustrating an example of a prompt related to the operation of the information processing system. FIG. 10 is a diagram illustrating an example of a response result related to the operation of the information processing system. FIG. 11 is a diagram illustrating an example of a message related to the operation of the information processing system.

[0019] The embodiments will be described in detail with reference to the drawings. However, the present invention is not limited to the following description, and it will be readily understood by those skilled in the art that various changes in form and details can be made without departing from the spirit and scope of the present invention. Therefore, the present invention should not be interpreted as being limited to the description of the embodiments shown below. In the configuration of the invention described below, the same parts or parts having similar functions will be denoted by the same reference numerals in different drawings, and repeated description thereof will be omitted.

[0020] In this specification, the ordinal numbers "first" and "second" are used for convenience and do not limit the number of components or the order of the components (e.g., the order of processes or the order of stacking). Furthermore, an ordinal number attached to a component in one part of this specification may not match an ordinal number attached to the same component in another part of this specification or in the claims. Furthermore, even if a term in this specification does not have an ordinal number attached to it, an ordinal number may be attached in the claims to avoid confusion between the components. Furthermore, even if a term in this specification has an ordinal number attached to it, the ordinal number may be omitted in the claims.

[0021] Embodiment 1 In this embodiment, an information processing system and an information processing method according to one embodiment of the present invention will be described with reference to FIGS.

[0022] FIG. 1 is a diagram illustrating an example of a configuration of an information processing system according to an embodiment of the present invention.

[0023] 1, the information processing system described in this embodiment includes an information processing device 10, an information processing device 40, and an information processing device 41. The information processing device 10, the information processing device 40, and the information processing device 41 are connected to each other via a network 30, and can transmit and receive text data and the like between them.

[0024] Furthermore, the information processing system described in this embodiment may be configured so that a user can input documents, etc. by directly operating the information processing device 10, or may be configured so that a user can input documents, etc. by using an information terminal 20 connected to the information processing device 10 via a network 31, as shown in Figure 1.

[0025] In the present embodiment, an example is shown in which the information processing system has three information processing devices (information processing device 10, information processing device 40, and information processing device 41), but the present invention is not limited to this. The information processing system may have one, two, or four or more information processing devices.

[0026] Configuration examples of the information processing device 10, the information processing device 40, the information processing device 41, the information terminal 20, the network 30, and the network 31 will be described below.

[0027] <<Configuration Example of Information Processing Device 10>> A configuration example of the information processing device 10 included in the information processing system of one embodiment of the present invention will be described with reference to FIG.

[0028] As shown in FIG. 2, the information processing device 10 includes an input unit 110 , a storage unit 120 , a processing unit 130 , an output unit 140 , and a transmission path 150 .

[0029] In the drawings accompanying this specification, the components are classified by function and shown as independent blocks in the block diagrams. However, in reality, it is difficult to completely separate the components by function, and one component may be involved in multiple functions. For example, part of the processing unit 130 may function as the input unit 110. Also, one function may be involved in multiple components. For example, the processing performed by the processing unit 130 may be executed by different servers depending on the processing.

[0030] [Input Unit 110] The input unit 110 can accept data from outside the information processing device 10. For example, the input unit 110 can accept data from the information terminal 20. The input unit 110 can also accept data from the information processing device 40 and the information processing device 41. Specifically, a device such as a personal computer having a communication port or a communication function can be used.

[0031] The input unit 110 can supply the received data to one or both of the storage unit 120 and the processing unit 130 via the transmission path 150 .

[0032] [Storage Unit 120] The storage unit 120 has a function of storing a program executed by the processing unit 130. The storage unit 120 may also have a function of storing data generated by the processing unit 130 (e.g., calculation results, analysis results, inference results), data accepted by the input unit 110, and the like.

[0033] The storage unit 120 may have a database. Furthermore, the information processing device 10 may have a database separate from the storage unit 120. The information processing device 10 may have a function to retrieve data from a database that exists outside the storage unit 120, outside the information processing device 10, or outside the information processing system. Furthermore, the information processing device 10 may have a function to retrieve data from both its own database and an external database.

[0034] Either or both of a storage and a file server can be used as the memory unit 120. Also, a database that records paths of files stored in a file server can be used as the memory unit 120.

[0035] The storage unit 120 includes at least one of a volatile memory and a nonvolatile memory. Examples of the volatile memory include a dynamic random access memory (DRAM) and a static random access memory (SRAM). Examples of the nonvolatile memory include a resistive random access memory (ReRAM), a phase change random access memory (PRAM), a ferroelectric random access memory (FeRAM), a magnetoresistive random access memory (MRAM), and a flash memory. The storage unit 120 may also include at least one of NOSRAM (registered trademark) and DOSRAM (registered trademark).

[0036] The storage unit 120 may include a recording media drive, such as a hard disk drive (HDD) or a solid state drive (SSD).

[0037] NOSRAM is an abbreviation for "Nonvolatile Oxide Semiconductor Random Access Memory." NOSRAM refers to a memory device in which a memory cell is a two-transistor (2T) or three-transistor (3T) gain cell, and the transistor included in the memory cell is a transistor using a metal oxide for a channel formation region (also referred to as an OS transistor).

[0038] An OS transistor has an extremely small leakage current, that is, a current that flows between the source and drain in an off state. NOSRAM can be used as a nonvolatile memory by retaining a charge corresponding to data in a memory cell using its extremely small leakage current characteristic. NOSRAM is particularly suitable for arithmetic processing in which only data read operations are repeated in large quantities because it can read stored data without destroying it (nondestructive read). Since NOSRAM can increase its data capacity by stacking layers, it can be used as a large-scale cache memory, main memory, or storage memory to improve the performance of semiconductor devices.

[0039] DRAM refers to a RAM (Random Access Memory) having 1T (transistor) 1C (capacitance) type memory cells. DOSRAM is an abbreviation for "Dynamic Oxide Semiconductor RAM." DOSRAM is a DRAM formed using OS transistors, and is a memory that temporarily stores information sent from an external device. DOSRAM utilizes the low off-state current of OS transistors, which can suppress data degradation due to the off-state current and retain data for a long period of time. This also reduces the number of times data needs to be refreshed, so using DOSRAM can reduce power consumption.

[0040] In this specification and the like, a metal oxide refers to an oxide of a metal in a broad sense. Metal oxides are classified into oxide insulators, oxide conductors (including transparent oxide conductors), oxide semiconductors (also referred to as oxide semiconductors or simply as OSs), and the like. For example, when a metal oxide is used for a semiconductor layer of a transistor, the metal oxide may be referred to as an oxide semiconductor.

[0041] The oxide semiconductor, which is a type of metal oxide, preferably contains indium (In). For example, indium oxide (InO ｘ, x is a real number greater than 0). When the oxide semiconductor is a metal oxide containing indium, the carrier mobility (electron mobility) of the OS transistor is increased. Furthermore, the oxide semiconductor preferably contains an element M. The element M is preferably at least one of aluminum (Al), gallium (Ga), and tin (Sn). Other elements applicable to the element M include boron (B), silicon (Si), titanium (Ti), iron (Fe), nickel (Ni), germanium (Ge), yttrium (Y), zirconium (Zr), molybdenum (Mo), lanthanum (La), cerium (Ce), neodymium (Nd), hafnium (Hf), tantalum (Ta), and tungsten (W). However, a combination of two or more of the above elements may be used as the element M. The element M is, for example, an element having a high bond energy with oxygen. For example, the element M is an element having a higher bond energy with oxygen than indium. The metal oxide contained in the channel formation region is preferably a metal oxide containing zinc (Zn), since zinc-containing metal oxides may be easily crystallized.

[0042] The oxide semiconductor used for a semiconductor layer including a channel formation region of a transistor is not limited to a metal oxide containing indium. For example, a metal oxide containing zinc but not indium, such as zinc tin oxide or gallium tin oxide, a metal oxide containing gallium, or a metal oxide containing tin can be used as the oxide semiconductor.

[0043] [Processing Unit 130] The processing unit 130 has a function of performing processes such as calculation, analysis, and inference using data supplied from one or both of the input unit 110 and the storage unit 120. The processing unit 130 can supply generated data (e.g., calculation results, analysis results, and inference results) to one or both of the storage unit 120 and the output unit 140.

[0044] The processing unit 130 has a function of acquiring data from the storage unit 120. The processing unit 130 may also have a function of recording or registering data in the storage unit 120.

[0045] The processing unit 130 may include, for example, an arithmetic circuit, a central processing unit (CPU), and a graphics processing unit (GPU).

[0046] The processing unit 130 may include a microprocessor such as a DSP (Digital Signal Processor). The microprocessor may be implemented by a PLD (Programmable Logic Device) such as an FPGA (Field Programmable Gate Array) or an FPAA (Field Programmable Analog Array). The processing unit 130 may also include a quantum processor. The processing unit 130 can perform various data processing and program control by interpreting and executing instructions from various programs using the processor. Programs that can be executed by the processor are stored in at least one of the memory area of ​​the processor and the storage unit 120.

[0047] The processing unit 130 may include a main memory. The main memory may include at least one of a volatile memory such as a RAM and a non-volatile memory such as a ROM (Read Only Memory). The main memory may also include at least one of the NOSRAM and DOSRAM.

[0048] The RAM may be, for example, a DRAM or an SRAM, and a virtual memory space is allocated and used as a working space for the processing unit 130. The operating system, application programs, program modules, program data, lookup tables, and the like stored in the storage unit 120 are loaded into the RAM for execution. The data, programs, and program modules loaded into the RAM are each directly accessed and operated by the processing unit 130.

[0049] The ROM can store a BIOS (Basic Input / Output System), firmware, etc., which do not require rewriting. Examples of ROM include mask ROM, OTPROM (One Time Programmable Read Only Memory), and EPROM (Erasable Programmable Read Only Memory). Examples of EPROMs include UV-EPROMs (Ultra-Violet Erasable Programmable Read Only Memories), which allow stored data to be erased by exposure to ultraviolet light, EEPROMs (Electrically Erasable Programmable Read Only Memories), and flash memories.

[0050] The processing section 130 can include one or both of an OS transistor and a transistor having silicon in a channel formation region (Si transistor).

[0051] The processing unit 130 preferably includes an OS transistor. Because an OS transistor has an extremely low off-state current, using the OS transistor as a switch for retaining charge (data) flowing into a capacitor functioning as a memory element can ensure a long data retention period. By utilizing this characteristic in at least one of the register and cache memory of the processing unit, the processing unit can be operated only when necessary and can be turned off at other times by saving information from the previous processing to the memory element. In other words, normally-off computing is possible, thereby enabling low power consumption of the information processing system. Note that when the processing unit 130 includes a central processing unit (CPU) and the CPU performs the above-described operations, the CPU may be referred to as a NoffCPU (registered trademark).

[0052] It is preferable that the information processing device 10 uses artificial intelligence (AI) for at least part of the processing.

[0053] In particular, it is preferable that a neural network be used in the information processing device 10. The neural network is realized by a circuit (hardware) or a program (software).

[0054] In this specification, a neural network refers to a general model that mimics the neural circuit network of a living organism, determines the connection strength between neurons through learning, and has problem-solving capabilities. A neural network has an input layer, an intermediate layer (hidden layer), and an output layer.

[0055] In this specification and the like, when discussing neural networks, determining the connection strengths (also called weighting coefficients) between neurons from existing information may be referred to as "learning."

[0056] In this specification and the like, the act of constructing a neural network using connection strengths obtained by learning and deriving a new conclusion from it may be referred to as "inference."

[0057] The information processing device 10 can perform processing using a natural language processing model that uses AI. For example, the information processing device 10 can perform processing using a natural language processing model such as BERT (Bidirectional Encoder Representations from Transformers) or T5 (Text-to-Text Transfer Transformer).

[0058] The information processing device 10 can also use a morphological analyzer, which is a natural language processing expression learning technology. For example, processing can be performed using a morphological analyzer such as "Mecab" or "Sudachi." Specifically, even in a language such as Japanese where there is no clear division between consecutive words, words can be divided. The part of speech of a word can also be obtained. An appropriate part of speech, such as a noun or adjective, can be selected. An appropriate phrase can also be selected by combining consecutive words.

[0059] [Output Unit 140] The output unit 140 can output at least one of the calculation result, analysis result, and inference result in the processing unit 130 to the outside of the information processing device 10.

[0060] For example, the output unit 140 can transmit data to the information terminal 20, the information processing device 40, and the information processing device 41. As the output unit 140, for example, a device such as a wired communication port, a wireless communication port, or an optical communication port can be used.

[0061] [Transmission Path 150] The transmission path 150 has a function of transmitting data. Data can be transmitted and received between the input unit 110, the storage unit 120, the processing unit 130, and the output unit 140 via the transmission path 150. The transmission path 150 may be, for example, a bus line on a motherboard, a wired communication cable, or an optical communication cable.

[0062] <<Configuration Example of Information Processing Device 40>> The information processing device 40 can process received data and transmit the processing results. For example, the information processing device 40 can perform processing such as calculations using data received from the information processing device 10. Furthermore, the information processing device 40 can transmit the processing results to the information processing device 10. This can reduce the calculation load on the information processing device 10.

[0063] The information processing device 40 can perform processing using a natural language processing model that uses AI. For example, the information processing device 40 can perform processing using an embedding model, which is a natural language processing technology that uses AI. As a natural language processing model that performs processing of the embedding model, the information processing device 40 can execute processing using a natural language processing model such as BERT (Bidirectional Encoder Representations from Transformers) or Word2vec (Word to vector).

[0064] An embedding model is a natural language processing technique for converting language data into embedded representations. Embedded representations convert language data, such as text or words, into quantified vectors that can be used in applications such as search.

[0065] The information processing device 40 is a large-scale computer such as a server computer or a supercomputer. Preferably, the information processing device 40 has a function as a parallel computer. By using the information processing device 40 as a parallel computer, it is possible to perform large-scale calculations required for AI learning and inference, for example.

[0066] The information processing device 40 is a computer with higher processing power than the information processing device 10. For example, if both the information processing device 10 and the information processing device 40 have functions as parallel computers, the information processing device 40 has higher processing power than the information processing device 10 and can perform large-scale calculations. Also, for example, if both the information processing device 10 and the information processing device 40 can perform processing using a model that utilizes a large-scale language model, the information processing device 40 can execute processing using a large-scale AI model compared to the information processing device 10.

[0067] It should be noted that the service provider does not necessarily have to own the information processing device 40. For example, the service provider can use part of the service provided by another business or the like using the information processing device 40.

[0068] <<Configuration Example of Information Processing Device 41>> The information processing device 41 can process received data and transmit the processing results. For example, the information processing device 41 can perform processing such as calculations using data received from the information processing device 10. The information processing device 41 can also transmit the processing results to the information processing device 10. This can reduce the calculation load on the information processing device 10.

[0069] The information processing device 41 can perform processing using a natural language processing model that uses AI. For example, the information processing device 41 can perform processing using a natural language processing model such as BERT (Bidirectional Encoder Representations from Transformers) or T5 (Text-to-Text Transfer Transformer).

[0070] Furthermore, the information processing device 41 can perform processing using a model that utilizes a large-scale language model (such as a sentence generation model or a dialogue model). For example, the information processing device 41 can perform processing using a large-scale language model such as GPT-3 (registered trademark), GPT-3.5, GPT-4 (registered trademark), LaMDA (Language Model for Dialogue Applications), PaLM (Pathways Language Model), or Llama2.

[0071] Furthermore, the information processing device 41 can execute processing using a general-purpose language processing model that can perform various natural language processing tasks.

[0072] The information processing device 41 is a large computer such as a server computer or a supercomputer. Preferably, the information processing device 41 has a function as a parallel computer. By using the information processing device 41 as a parallel computer, it is possible to perform large-scale calculations required for AI learning and inference, for example.

[0073] It is preferable that the information processing device 41 is a computer with higher processing power than the information processing device 10. For example, if both the information processing device 10 and the information processing device 41 have functions as parallel computers, it is preferable that the information processing device 41 has higher processing power than the information processing device 10 and can perform large-scale calculations. Also, for example, if both the information processing device 10 and the information processing device 41 can perform processing using a model that utilizes a large-scale language model, it is preferable that the information processing device 41 executes processing using a large-scale AI model than the information processing device 10.

[0074] It should be noted that the service provider does not necessarily have to own the information processing device 41. For example, the service provider can use part of the service provided by another business or the like using the information processing device 41.

[0075] As described above, the information processing system of one embodiment of the present invention may include the information processing device 10 that performs processing using a morphological analyzer, the information processing device 40 that performs processing using a natural language processing model, and the information processing device 41 that performs processing using a large-scale language model. Alternatively, the information processing system of one embodiment of the present invention may be configured to perform both processing using the morphological analyzer and processing using the natural language processing model using a single computer. In other words, the information processing device 40 may have the functions of the above-described information processing device 10. Alternatively, the information processing system of one embodiment of the present invention may be configured to perform all of processing using the morphological analyzer, processing using the natural language processing model, and processing using the large-scale language model using a single computer. In other words, the information processing device 41 may have the functions of the above-described information processing device 10 and the functions of the information processing device 40.

[0076] <<Configuration Example of Information Terminal 20>> The information terminal 20 can accept data input by a user of the information processing system of one aspect of the present invention. The information terminal 20 can also provide the user with data output by the information processing system of one aspect of the present invention.

[0077] Furthermore, the information terminal 20 can transmit data received from the user to the information processing device 10. Furthermore, the information terminal 20 can provide data received from the information processing device 10 to the user.

[0078] Furthermore, the information terminal 20 can convert data received from a user into data that can be handled by the information processing device 10, and can transmit the data to the information processing device 10. Furthermore, the information terminal 20 can convert data received from the information processing device 10 into data that can be viewed by the user, and provide the data to the user.

[0079] For example, dedicated application software or a web browser is installed on the information terminal 20. A user can access the information processing device 10 via either of these. This allows the user to enjoy services using the information processing system according to one aspect of the present invention, for example, using a computer with lower processing power than the information processing device 10.

[0080] The information terminal 20 can also be called a client computer, etc. Each of the information terminals 20 is an information terminal device used by a user of the information processing system according to one aspect of the present invention.

[0081] For example, a desktop computer 20a, a notebook computer 20b, a smartphone 20c, or a tablet computer 20d can be used as the information terminal 20. The tablet computer 20d can also be used as a notebook computer by connecting it to a housing 21 having a keyboard.

[0082] This allows a user of the information processing system to use, for example, a computer such as information terminal 20 that has lower processing power than information processing device 10, information processing device 40, or information processing device 41 to control an information processing system of one embodiment of the present invention, give instructions to information processing device 40 or information processing device 41, and enjoy services.

[0083] <<Configuration Example of Network 30 and Network 31>> When performing wired communication in one or both of the networks 30 and 31, specifications standardized by IEEE such as Ethernet (registered trademark) can be used. In addition, the type of communication can be electrical communication using electric wires such as twisted pair cables, optical communication using optical fibers, etc.

[0084] On the other hand, when wireless communication is performed as one or both of network 30 and network 31, communication standards such as the fourth generation mobile communication system (4G), fifth generation mobile communication system (5G), and sixth generation mobile communication system (6G), or specifications standardized by IEEE such as Wi-Fi (registered trademark) and Bluetooth (registered trademark), can be used as communication protocols or communication technologies.

[0085] One or both of the networks 30 and 31 may be, for example, a PAN (Personal Area Network), a LAN (Local Area Network), a CAN (Campus Area Network), a MAN (Metropolitan Area Network), a WAN (Wide Area Network), a GAN (Global Area Network), etc.

[0086] For example, by using a GAN for one or both of the networks 30 and 31, it is possible to use the Internet, which is the foundation of the World Wide Web (WWW).

[0087] Furthermore, a person who provides a service using the information processing system of one embodiment of the present invention can provide the service using the information processing system of one embodiment of the present invention via the network 31, for example.

[0088] In an information processing system according to one embodiment of the present invention, when at least one of the networks 30 and 31 is constructed as a LAN, the possibility of confidential information leaks can be reduced, for example, compared to when the Internet is used.

[0089] Next, an example of an information processing method used in the information processing system described in this embodiment will be described with reference to FIGS.

[0090] 3 and 4 are flowcharts illustrating an example of an information processing method according to one embodiment of the present invention. As shown in FIG. 3, the information processing method includes steps S101 to S112.

[0091] 5 shows each process included in the information processing method, which is supported by the information terminal 20, the information processing device 10, the information processing device 40, and the information processing device 41. Each process will be described below.

[0092] When an information processing method according to an embodiment of the present invention is "started," step S101 in Fig. 3 is first performed. In step S101, the information processing device 10 acquires pairs including a first sentence written in a first language and a second sentence obtained by translating the first sentence into a second language. The pairs are input to the information processing device 10 from the information terminal 20 by, for example, a user. The number of pairs may be one or more.

[0093] The first sentence can also be referred to as an original sentence. The second sentence can also be referred to as a translation. The first language and the second language are not particularly limited, and different languages ​​can be selected from, for example, Japanese, English, German, French, Chinese, Korean, Spanish, Portuguese, Russian, Arabic, Hindi, etc. In this embodiment, an example will be described in which Japanese is used as the first language and English is used as the second language.

[0094] 3, the information processing device 10 determines whether there are any unprocessed pairs among the pairs received in step S101. If there are any unprocessed pairs, the process proceeds to step S103. If all of the acquired pairs have been processed, the process proceeds to step S112.

[0095] <Example of operation when there are unprocessed pairs> In step S103 of FIG. 3, the information processing device 10 extracts one pair acquired in step S101, performs morphological analysis on the first sentence included in the pair, and divides the first sentence into multiple first phrases.

[0096] In step S103 of FIG. 3, the information processing device 10 performs morphological analysis on the second sentence included in the pair, and divides the second sentence into a plurality of second phrases.

[0097] Furthermore, multiple phrases can be obtained by performing morphological analysis on a sentence and combining words based on part-of-speech rules. For example, a sentence can be divided into multiple phrases by combining nouns, combining adjectives and nouns, or excluding particles.

[0098] In step S104 of FIG. 3, the information processing device 40 converts each of the plurality of first phrases and the plurality of second phrases into a vector.

[0099] Details of the process of step S104 in Fig. 3 will be described with reference to Fig. 4A. As shown in Fig. 4A, step S104 includes, for example, steps S104a to S104c.

[0100] In step S104a of FIG. 4A, the information processing device 10 transmits a plurality of first phrases and a plurality of second phrases to the information processing device 40.

[0101] In step S104b of FIG. 4A, the information processing device 40 converts a plurality of first phrases into a plurality of first vectors and a plurality of second phrases into a plurality of second vectors using an embedding model.

[0102] In step S104c of FIG. 4A, the information processing device 40 transmits the first vector and the second vector obtained in step S104b to the information processing device 10.

[0103] 3, the information processing device 10 calculates the similarity from the first vector and the second vector obtained in step S104. Furthermore, a similarity matrix is ​​created using the calculated similarity.

[0104] For one first vector, the similarity to each of the second vectors is calculated, and for one second vector, the similarity to each of the first vectors is calculated. Note that in this information processing method, when calculating the similarity, calculation is not performed between first vectors or between second vectors.

[0105] Methods for calculating the similarity between two vectors include cosine similarity, covariance, unbiased covariance, and Pearson's product-moment correlation coefficient. Among these, when evaluating the similarity between two sentences, it is preferable to use cosine similarity.

[0106] Cosine similarity can be calculated simply by taking the dot product of two vectors. Cosine similarity is expressed as a real number between -1 and 1, and the closer it is to 1, the higher the similarity. Therefore, by using cosine similarity, the degree of similarity between two sentences can be easily evaluated.

[0107] FIG. 6A shows an example in which the similarity between a first vector and a second vector is calculated using cosine similarity, and a similarity matrix between a first phrase and a second phrase is created.

[0108] The first phrases correspond to, for example, J1 to J5 in Fig. 6A and are listed vertically (Y-axis direction), and the second phrases correspond to, for example, E1 to E5 in Fig. 6A and are listed horizontally (X-axis direction).

[0109] In addition, in FIGS. 6A to 8B, as an example, J1 to J5 indicate Japanese phrases, and E1 to E5 indicate English phrases.

[0110] In step S106 in FIG. 3, the information processing device 10 normalizes the similarity matrix shown in FIG. 6A to create a normalized matrix.

[0111] The reason for normalizing the data is that the first phrases (J1 to J5) and the second phrases (E1 to E5) are written in different languages, and the similarity values ​​vary slightly. Normalizing the data makes it possible to adjust the scale of the data to a common standard, facilitating comparison and analysis.

[0112] The method of normalizing the similarity matrix will be described below. First, normalization in the X-axis direction is performed using the similarity matrix of Fig. 6A, and similarly, normalization in the Y-axis direction is performed using the similarity matrix of Fig. 6A.

[0113] Normalization in the X-axis direction refers to a process of converting the similarity in the X-axis direction shown in Fig. 6A into a real number between 0 and 1. Note that here, the Min-Max method is used as the method for normalization in the X-axis direction. Specifically, in each row of Fig. 6A, the Min-Max method is used to convert each value into a real number between 0 and 1, inclusive, thereby normalizing the matrix in Fig. 6A. In this specification, a matrix normalized in this way is referred to as an X-normalized matrix.

[0114] FIG. 6B is an example of an X normalized matrix created by the above method.

[0115] Normalization in the Y-axis direction refers to a process of converting the similarity in the Y-axis direction shown in Fig. 6A into a real number between 0 and 1. Note that here, as with Fig. 6B, the Min-Max method is used as the method for normalizing the Y-axis direction in Fig. 6A. Specifically, in each column of Fig. 6A, the Min-Max method is used to convert each value into a real number between 0 and 1, inclusive, thereby normalizing the matrix in Fig. 6A. In this specification, a matrix normalized in this manner is referred to as a Y-normalized matrix.

[0116] FIG. 6C is an example of a Y normalized matrix created by the above method.

[0117] Then, the similarity matrix in Fig. 6A can be normalized by multiplying the X normalization matrix by the Y normalization matrix. This matrix is ​​also called a normalization matrix.

[0118] FIG. 6D is an example of a normalized matrix created by the above method.

[0119] In step S107 in FIG. 3, the information processing device 10 randomizes the normalized matrix shown in FIG. 6D and creates a randomized matrix.

[0120] Specifically, by using the normalized matrix shown in Figure 6D, the sum of each row is calculated, and then the value of each row is divided by this sum, thereby randomizing the normalized matrix in the X-axis direction. The result of dividing the sum by the value of each row is the probability. This matrix is ​​also called the X probability matrix.

[0121] FIG. 7A is an example of an X probability matrix created by the above method.

[0122] Similarly, by using the normalized matrix shown in Figure 6D to calculate the sum of each column and then dividing the value of each column by that sum, the normalized matrix can be randomized along the Y axis. The result of dividing the sum by the value of each column is the probability. This matrix is ​​called the Y probability matrix.

[0123] FIG. 7B is an example of a Y probability matrix created by the above method.

[0124] The method for creating the probability matrix is ​​not limited to the above. For example, a probability matrix can be created based on the normalized matrix of Fig. 6D using a softmax function.

[0125] In step S108 of FIG. 3, the information processing device 10 refers to the X probability matrix shown in FIG. 7A and the Y probability matrix shown in FIG. 7B, obtains all matrix elements whose probability is higher than a threshold, and creates a correspondence table listing combinations with the first phrase and the second phrase.

[0126] When obtaining matrix elements, a predetermined threshold must be set, and the threshold can be set arbitrarily by the user. However, in this specification, as an example, the threshold is set to 0.3.

[0127] For example, referring to the X probability matrix shown in FIG. 7A, combinations of the first phrase and the second phrase that have a probability higher than the threshold value of 0.3 include the combination of "J1" and "E5", the combination of "J2" and "E2", the combination of "J2" and "E4", the combination of "J3" and "E2", the combination of "J4" and "E4", the combination of "J5" and "E2", and the combination of "J5" and "E4".

[0128] Also, for example, referring to the Y probability matrix shown in Figure 7B, combinations of the first phrase and the second phrase that have a probability higher than the threshold value of 0.3 include the combination of "J1" and "E5", the combination of "J3" and "E1", the combination of "J3" and "E2", the combination of "J4" and "E3", and the combination of "J4" and "E4".

[0129] From the above, if we extract only combinations of the first phrase and the second phrase that have probabilities higher than the threshold value of 0.3 in the X probability matrix of Figure 7A and the Y probability matrix of Figure 7B, we get the combinations "J1" and "E5," "J3" and "E2," and "J4" and "E4." These combinations are entered in the correspondence table.

[0130] In FIG. 8A, a first list 50 is an example of a correspondence table.

[0131] In this way, by extracting combinations with probabilities equal to or greater than a threshold from each of the X probability matrix and the Y probability matrix, it is possible to create a correspondence table between the first phrase and the second phrase.

[0132] In step S109 of FIG. 3, the information processing device 10 acquires the first phrases and the second phrases that are not included in the correspondence table created in step S108, and creates an unmatched list.

[0133] Specifically, the first phrases not included in the correspondence table of Figure 8A are "J2" and "J5." Also, the second phrases not included in the correspondence table of Figure 8A are "E1" and "E3." Therefore, the unmatched phrase list created in step S108 is, for example, as shown in Figure 8B.

[0134] 8B, the second list 51 is an example of an unmatched list for the first phrase, and the third list 52 is an example of an unmatched list for the second phrase.

[0135] Here, we will explain the case where the probability of each combination of a second phrase and two or more first phrases is equal to or greater than a threshold. For example, in Figures 7A and 7B, the first phrase "J5" is not associated with any of the multiple second phrases. However, if the probability of the combination of "J5" and "E4" in each of the X probability matrix and the Y probability matrix becomes higher than a threshold, both "J4" and "J5" will be listed as first phrases to be combined with "E4." In this case, since a unique first phrase cannot be associated with "E4," "J5" and "E4" are added to the unmatched list. Furthermore, "J4" and "E4," which are correctly associated, are also discarded from the correspondence table, and "J4" is added to the unmatched list. The above also applies to the case where the probability of each combination of a first phrase and two or more second phrases is higher than a threshold.

[0136] 3, operations related to prompts for input to a large-scale language model are performed, such as creating a prompt, sending a prompt, and receiving a response corresponding to the prompt.

[0137] The process of step S110 in Fig. 3 will be described in detail with reference to Fig. 4B. The prompt will be described in detail with reference to Fig. 9. The response result to the prompt will be described in detail with reference to Fig. 10.

[0138] As shown in FIG. 4B, step S110 includes steps S110a to S110d, for example.

[0139] In step S110a of Fig. 4B, a prompt including an instruction sentence, a first sentence, a second sentence, a correspondence table, and an unsupported list is created. Note that the prompt can be created, for example, by inserting the information included in the first sentence, the second sentence, the correspondence table, and the unsupported list into the template shown in Fig. 9.

[0140] 9 shows an example of the prompt, a prompt 60. The prompt 60 includes a first item 61, a second item 62, a third item 63, a first area 64, and a second area 65.

[0141] 9, a first item 61 includes a directive. The directive includes content that, for a first phrase that is not associated with a second phrase, causes a corresponding second phrase to be selected from a plurality of second phrases acquired from the unassociated list. The directive also includes content that causes a reason for selecting the second phrase to be indicated. The directive also includes content that causes a reason for pairing for each combination of a first phrase and a second phrase to be indicated.

[0142] Specifically, the first item 61 includes, as an example, an instruction such as, "List pairs of corresponding phrases in the first sentence and the second sentence shown below. Select the appropriate phrase from the first area 64 for <?en> in the alignment list. Also, enter the reason for selecting <?en> in <detailed reason>."

[0143] In FIG. 9, the second item 62 includes the first sentence (original sentence).

[0144] In FIG. 9, the third item 63 includes the second sentence (translation).

[0145] In FIG. 9, the first area 64 lists the second phrases obtained from the unmatched list created in step S109.

[0146] 9, the second region 65 includes an alignment list. The alignment list lists paired combinations of first phrases and second phrases based on the correspondence table and the unmatched list. The second region 65 includes second region 65a, second region 65b, second region 65c, and second region 65d.

[0147] The second area 65a lists a plurality of first phrases obtained from the correspondence table and the uncorrespondence list.

[0148] The second region 65b displays a symbol indicating that the first phrase and the second phrase are a pair. Note that, although the symbol is displayed as "=" in Fig. 9, one embodiment of the present invention is not limited to this, and characters other than "=" may be used as the symbol.

[0149] The second area 65c lists a plurality of second phrases obtained from the correspondence table. In the second area 65c, the first phrase obtained from the unmatched list is not associated with a second phrase. If there is no corresponding second phrase, "<?en>" is entered in the second area 65c as a fixed phrase.

[0150] The second area 65d displays the fixed phrase "(<detailed reason>)."

[0151] In step S110b of FIG. 4B, the prompt created in step S110a is sent to the information processing device 41.

[0152] In step S110c of FIG. 4B, the information processing device 41 uses a large-scale language model to generate a response to the prompt.

[0153] 10 shows a report 70 as an example of a response to the prompt. The report 70 includes a third area 71.

[0154] 10, the third area 71 shows an alignment list, and displays the first phrases, the second phrases, and the reason for the determination included in the prompt. The third area 71 includes a third area 71a, a third area 71b, a third area 71c, and a third area 71d.

[0155] The third area 71a lists a plurality of first phrases obtained from the prompt. It is preferable that the contents of the second area 65a and the third area 71a are the same.

[0156] The third area 71b displays a symbol indicating that the first phrase and the second phrase are a pair. Note that, although the symbol is displayed as ":" in Fig. 10, one embodiment of the present invention is not limited to this, and a character other than ":" may be used as the symbol.

[0157] The third area 71c lists a plurality of second phrases obtained from the prompt.

[0158] In the third area 71c, the fixed phrase indicated as "<?en>" in the information obtained from the prompt is replaced with a second phrase selected using a large-scale language model. Specifically, a plurality of second phrase candidates are listed in the first area 64 of Fig. 9, and an appropriate phrase is selected from among them to replace the fixed phrase.

[0159] The third area 71d shows the reason why the first phrase and the second phrase are paired, which is generated using a large-scale language model.

[0160] In step S110d of FIG. 4B, the response result generated using the large-scale language model of the information processing device 41 is received by the information processing device 10.

[0161] In step S111 of FIG. 3, the information processing device 10 creates a bilingual dictionary that stores a plurality of first phrases, a plurality of second phrases, and reasons for judgment based on the response result.

[0162] In an information processing method according to one aspect of the present invention, instead of simply inputting the original text and its translation into a large-scale language model, a correspondence table of words in the original text and its translation is created in advance and input to the large-scale language model. This increases the accuracy of the answers provided by the large-scale language model. Furthermore, even if there are unmatched parts in the correspondence table, inputting them into the large-scale language model makes it possible to supplement the information and obtain the reasons for the correspondence.

[0163] Returning to step S102 in FIG. 3, if there are unprocessed pairs, steps S103 to S111 in FIG. 3 are repeatedly performed until there are no unprocessed pairs.

[0164] <Example of Operation When There Are No Unprocessed Pairs> When there are no unprocessed pairs, in step S112 of FIG. 3, the information processing device 10 refers to the bilingual dictionary and outputs a warning regarding translation variations.

[0165] Specifically, the details of the process of step S112 in Fig. 3 will be described with reference to Fig. 4C. As shown in Fig. 4C, step S112 includes, for example, step S112a and step S112b.

[0166] In step S112a of FIG. 4C, the bilingual dictionary is referenced, and if there are multiple second phrase candidates for one first phrase, the multiple second phrases and the reason for the determination are output to the information terminal 20.

[0167] In step S112b, the user can check the second phrases and the reason for the judgment received by the information terminal 20.

[0168] 11 shows a message 80 as an example of information displayed on the information terminal 20. The message 80 includes a fourth item 81, a fifth item 82, and a fourth area 83.

[0169] The fourth item 81 lists the first phrase.

[0170] The fifth item 82 lists a plurality of second phrases.

[0171] The fourth area 83 lists the reasons for the determination for each of the plurality of second phrases. The user can check the reasons for the determination and select an appropriate translation for the first phrase in the fourth item 81 from the second phrases in the fifth item 82. It is preferable that the information processing device 10 accepts the user's selection of a translation and updates the bilingual dictionary.

[0172] In step S112b of FIG. 4C, if the information terminal 20 receives a warning about translation variations, the information processing method according to one aspect of the present invention can be "ended."

[0173] Preferably, the information processing device 10 outputs the report 70. An example of information displayed on the information terminal 20 is the report 70 shown in FIG. 10 . Preferably, the information terminal 20 displays, for example, a warning about translation variations and the report 70. Furthermore, if there is no warning about translation variations, the information terminal 20 can "end" the information processing method of one aspect of the present invention by receiving either or both of a report indicating that there is no warning and the report 70.

[0174] By checking the notes regarding translation variations, the user can check phrases that are candidates for translation along with the reasons for the variations. This makes it easy to check and correct translation variations. Furthermore, by checking the report 70, the user can easily check the correspondence between phrases in the original text and the translation. Furthermore, the correspondence between phrases in the original text and the translation can be checked along with the reasons for the correspondence. This simplifies and streamlines the process of checking translated sentences.

[0175] By using the information processing system described in the above embodiment, it is possible to easily check the translated text. Also, if an error occurs during translation, support for correction can be received. Furthermore, it is possible to easily register unregistered words in a bilingual dictionary, and a bilingual dictionary can be easily constructed.

[0176] 10: information processing device, 20: information terminal, 20a: desktop computer, 20b: notebook computer, 20c: smartphone, 20d: tablet computer, 21: housing, 30: network, 31: network, 40: information processing device, 41: information processing device, 50: first list, 51: second list, 52: third list, 60: prompt, 61: first item, 62: second item, 63: third item, 64: first area, 65: second area, 70: report, 71: third area, 80: message, 81: fourth item, 82: fifth item, 83: fourth area, 110: input unit, 120: memory unit, 130: processing unit, 140: output unit, 150: transmission path

Claims

a first information processing device; a second information processing device; the first information processing device has first to eighth functions, the first function is a function of acquiring a first sentence written in a first language and a second sentence obtained by translating the first sentence into a second language; the second function is a function of dividing the first sentence into a plurality of first phrases; the third function is a function of dividing the second sentence into a plurality of second phrases; the fourth function is a function of calculating the probability that one of the plurality of first phrases and one of the plurality of second phrases are translations of each other, and creating a matrix; the fifth function is a function of creating a correspondence table in which combinations of the first phrase and the second phrase whose probability is higher than the threshold are listed, and an unmatched list in which phrases not included in the correspondence table are listed, the sixth function is a function of creating a prompt including the first sentence, the second sentence, the correspondence table, and the uncorrespondence list; the seventh function is a function of transmitting the prompt to the second information processing device; the eighth function is a function of acquiring, as a response result to the prompt, a phrase included in the unsupported list, a phrase determined to be a translation of the phrase, and a reason for the determination, from the second information processing device; The second information processing device a function of receiving the prompt and generating the response result using a large-scale language model; Information processing system. 　 a first information processing device; a second information processing device; a third information processing device; the first information processing device has first to twelfth functions, the first function is a function of acquiring a first sentence written in a first language and a second sentence obtained by translating the first sentence into a second language; the second function is a function of dividing the first sentence into a plurality of first phrases; the third function is a function of dividing the second sentence into a plurality of second phrases; the ninth function is a function of transmitting the plurality of first phrases and the plurality of second phrases to the third information processing device; the tenth function is a function of acquiring, from the third information processing device, a plurality of first vectors corresponding to the plurality of first phrases and a plurality of second vectors corresponding to the plurality of second phrases; the eleventh function is a function of calculating a similarity between the first vector and the second vector and creating a similarity matrix including the similarity; the twelfth function is a function of normalizing the similarity matrix to create a normalized matrix; the fourth function is a function of calculating, from the normalized matrix, a probability that one of the plurality of first phrases and one of the plurality of second phrases are translations of each other, and creating a matrix; the fifth function is a function of creating a correspondence table in which combinations of the first phrase and the second phrase whose probability is higher than the threshold are listed, and an unmatched list in which phrases not included in the correspondence table are listed, the sixth function is a function of creating a prompt including the first sentence, the second sentence, the correspondence table, and the uncorrespondence list; the seventh function is a function of transmitting the prompt to the second information processing device; the eighth function is a function of acquiring, as a response result to the prompt, a phrase included in the unsupported list, a phrase determined to be a translation of the phrase, and a reason for the determination, from the second information processing device; The second information processing device a function of receiving the prompt and generating the response result using a large-scale language model; The third information processing device a function of converting the first phrase into the first vector and converting the second phrase into the second vector by processing using an embedding model; Information processing system. 　 In claim 1 or claim 2, In the sixth function, the first sentence, the second sentence, the correspondence table, and the uncorresponding list are inserted into a template to create the prompt. Information processing system. In claim 1 or claim 2, The first information processing device a function of creating a bilingual dictionary that stores phrases included in the unsupported list, phrases determined to be translations of the phrases, and reasons for the determination, based on the response result; Information processing system. 　 In claim 4, The first information processing device a function of referring to the bilingual dictionary, and when there are a plurality of second phrases that are candidates for translation for one of the plurality of first phrases, outputting the candidate second phrases and the reason for the determination; Information processing system.

Images